Mechanical Beam Steering Using Gimbals and Fast Steering Mirrors
The simplest method of steering an optical beam is by use of mechanical means. An example of such standard mechanical t tans includes a standard, macro-scale gimbal 10, such as shown in FIG. 1. A light source and lens mounted on the gimbal 10 enables steering in any direction provided that the gimbal allows for rotation about the two primary axes. Macro-scale gimbals 10 are generally heavy (e.g., ˜20 lbs.), require significant power (e.g., ˜48 W) and are slow (e.g., <10 Hz); although smaller gimbals with improved performance are available, they are still macro-scale devices. For some applications (e.g. on board ships or fixed building installations), the Size, Weight, & Power (“SWAP”) of macro-scale gimbals is not prohibitive. However, in other applications, the large SWAP is prohibitive and other non-mechanical beam steering solutions are needed.
While gimbals are generally used to coarsely point the FSO terminal, standard fast steering mirrors (“FSM”) or standard piezo-controlled (“PZ”) mirrors are often used for fine pointing of both the transmit and the receive beams. FSMs (e.g. Newport FSM-300 FSM) and PZs (e.g., CONEX-AG-M100D PZ) typically operate by controlling the beam reflection from a 1″ mirror. The FSM deflects the beam by ±3°, whereas a PZ deflects the beam by ±0.75°. Because one FSM or PZ is needed for both transmit and receive beams, two units and controllers are required in each interrogator. The weight of the FSM mirror is ˜1 lb. and weight of the PZ is only ˜85 g. Each needs additional controller electronics. For larger scale motions (e.g., 25 mrad), the FSM can operate at up to 50 Hz and, for very small (0.1 mrad) control loop motions, at up to 580 Hz. The PZ controlled pitch/yaw mount is generally regarded as a slower (<10 Hz) system but with absolute position encoding it has a very stable and reproducible pointing capability.
Chip-Scale Mechanical Beam Steering Using MEMS
Another method of steering an optical beam is by use of standard micro-electro-mechanical systems (“MEMS”). An example of a standard MEMS device for beam steering is a standard MEMS-based tip/tilt/piston micromirror 20, such as shown in FIG. 2. In general MEMS approaches simply implement mechanical beam steering similar to a gimbal, but at the micro-scale (i.e. characteristic lengths of hundreds of micrometers). MEMS mirrors are often preferred over lenses due to significant previous research and development by Texas Instruments (e.g., Texas Instruments' digital micromirror device (“DMD”) for movie projectors) and Lucent Technologies (e.g., Lucent's optical cross-connect switch for optical networks). By fabricating a micro-mirror on a tip/tilt/piston actuator, the mirror can be positioned to reflect any incident light in a desired direction. Limitations of MEMS micromirrors include the limited response time (typically, in the 10 microsecond to millisecond-range) and modest fill-factor (e.g. 30% fill-factor and 7 dB insertion loss) due to the complicated actuator design which requires motion along several axes. For large-angle beam steering, a high fill-factor is essential since the steering angle is a direct function of the mirror separation and the amount of optical power in a steered beam depends on the mirror size and fill factor. Many optical MEMS components also only function as switches with two stable states (e.g. Texas Instruments' DMD) and are therefore not suitable for beam steering which requires continuously variable devices.
Chip-Scale Non-Mechanical Beam Steering Using Liquid Crystals
Another method of steering an optical beam is by use of standard liquid crystals (“LCs”). LCs are materials that can change their refractive index upon application of an electric field. For nematic LCs, the time-averaged field needs to be zero; otherwise, the LC will experience permanent ion migration and damage. Initial applications of liquid crystals to chip-scale beam steering have relied on surface-normal configurations, such as the surface-normal liquid crystal 30 shown in FIG. 3. Light is passed perpendicularly through a liquid-crystal phase modulator array. By applying an appropriate phase across the chip the incident beam can be steered in any direction. Initial demonstrations have shown the potential of using liquid crystals for up to +/−5° beam steering angles at sub-second response times. More recently, steering angles of +/−40° have been achieved using gratings (with the drawback that these devices are highly polarization dependent). The temporal response can be improved, although the speed is generally limited with surface-normal approaches. The requirement for a zero time-averaged electric field also may place some limitations on LC's and their applications.
Chip-Scale Non-mechanical Beam Steering Using Liquid Crystal-Clad Waveguides
Another method of steering an optical beam is by use of a standard in-plane waveguide-based approach. In this method, the liquid crystal forms a top cladding in a thin core waveguide fabricated on a chip. The guided optical mode experiences a significant modal overlap with the liquid crystal top cladding. By applying a bias across the liquid crystal, the waveguide mode can attain a variable phase shift. In order to achieve beam steering along the wafer plane, the waveguide is terminated with a prism. A sawtooth electrode above the liquid crystal enables a variable phase front to be applied to the guided mode resulting in a variable beam steering along the θ-angle.
Steering along the -angle is achieved using a standard Ulrich coupler. In an Ulrich coupler, the waveguide core is made progressively thinner until light leaks out of the core at which point it is emitted through the substrate. The emission angle is governed by Snell's law and therefore depends on the effective index of the tapered waveguide section. The liquid crystal provides a controllable means for determining the mode effective index and hence the emission angle out of the chip. The switching speed is substantially improved over surface-normal approaches (such as shown in FIG. 3), although the fastest response time is still sub-millisecond. The manufacture of the liquid crystal based optical phased arrays also requires many custom fabrication processes that can make large-scale production challenging and cost-prohibitive.
Chip-Scale Non-Mechanical Beam Steering Using Silicon Photonics
In light of some of the challenges in the development of chip-scale optical phased arrays for beam steering, as discussed above, recent efforts have focused on using a silicon platform for chip-scale beam steering. Silicon platforms enable devices to be fabricated in existing foundries, generally older semiconductor electronics manufacturing facilities. This may lead to significant cost-savings with the potential for mass-production and large-scale adaptation in consumer electronics and other products.
The basic approach for standard silicon photonic optical phased arrays is to couple light into a single waveguide on a silicon chip, split the light into multiple waveguides, apply a variable phase shift to each waveguide, and then emit the light from each waveguide. The emitted light from the waveguides interfere so that in the far-field the emission looks like a focused beam that can be steered in any direction by varying the phase.
For example, a standard silicon photonic optical phased array approach uses sequential 1×2 splitters to achieve 16 waveguide channels. A standard triangular-shaped thin film heater above the waveguide provides a linear phase gradient so that, a single control signal varies the phase shift via thermo-optic heating and hence steers the beam along the θ-angle. Light is coupled out of the chip using gratings fabricated along the length of each waveguide. Steering along the ϕ-angle is achieved by taking advantage of the wavelength-dependent emission angle of the grating out-couplers. Although this approach enables two-dimensional beam steering, it requires significant wavelength tuning (up to Δλ=100 nm) which is undesirable for many applications since it requires a tunable laser. Thermo-optic tuning can also require substantial electrical power (10's of mW to a few Watts depending on the size of the array). Finally, the cascaded 1×2 splitters result in a large separation between the output waveguides so that the steering angle and fill factors are limited. In practice, attempting to bring the output waveguides closer together tends to introduce phase errors.
A conventional improvement on the above-mentioned basic approach includes using a non-linear spacing between the waveguide emitters to help suppress the sidelobes present in phased arrays. This improvement results in larger steering angles. Indeed, a 10°-steering angle was achieved in silicon from which the authors claim >30°-steering angle for a beam propagating in free space. The improvement, however, also used thermo-optic phase shifters requiring substantial power. In order to reduce the power requirement, thermo-optic phase shifters utilizing direct current injection into a silicon waveguide has been used. For a standard, large-scale optical phased array using direct current injection, light from a single input waveguide is split into M-rows and N-columns. Each element-MN (row-M and column-N) has a grating out-coupler and a thermo-optic phase shifter utilizing direct current injection. In this manner, arbitrary phase profiles can be achieved so that any image can be displayed. The drawback with this approach, however, is complexity in the control signals: every emitter element requires a control signal (i.e., N2-controls for an N×N array).
The prior art devices discussed above suffer from one or more deficiencies that make them unattractive in practice. For example, they are either large or heavy (e.g., gimbals), have a slow temporal response (e.g., MEMS and liquid crystals), require tunable lasers (e.g., silicon phased arrays with wavelength-steerable gratings), have significant power requirements (e.g., most thermo-optic based approaches), or require complex control signals of order N2 (e.g., direct current injection in large-scale silicon photonic phased arrays).